In-Domain African Languages Translation Using LLMs and Multi-armed Bandits
Not on the map. Why?
Abstract (arXiv)
Neural Machine Translation (NMT) systems face significant challenges when working with low-resource languages, particularly in domain adaptation tasks. These difficulties arise due to limited training data and suboptimal model generalization, As a result, selecting an optimal model for translation is crucial for achieving strong performance on in-domain data, particularly in scenarios where fine-tuning is not feasible or practical. In this paper, we investigate strategies for selecting the most suitable NMT model for a given domain using bandit-based algorithms, including Upper Confidence Bound, Linear UCB, Neural Linear Bandit, and Thompson Sampling. Our method effectively addresses the resource constraints by facilitating optimal model selection with high confidence. We evaluate the approach across three African languages and domains, demonstrating its robustness and effectiveness in both scenarios where target data is available and where it is absent.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 29 of its 51 references that are in the map. First posted on arXiv 21 May 2025.
Builds on
29 of 51 references.
References · 29
- 2024The Llama 3 Herd of Modelscited by 14,715
- 2018A Call for Clarity in Reporting BLEU Scorescited by 1,799
- 2022No Language Left Behind: Scaling Human-Centered Machine Translationcited by 1,312
- 2023LLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusioncited by 564
- 2024Aya Model: An Instruction Finetuned Open-Access Multilingual Language Modelcited by 295
- 2019Simple, Scalable Adaptation for Neural Machine Translationcited by 242
- 2022CometKiwi: IST-Unbabel 2022 Submission for the Quality Estimation Shared Taskcited by 207
- 2023MADLAD-400: A Multilingual And Document-Level Large Audited Datasetcited by 181
- 2019CCAligned: A Massive Collection of Cross-Lingual Web-Document Pairscited by 136
- 2021SimCLS: A Simple Framework for Contrastive Learning of Abstractive Summarizationcited by 132
- 2021Neural Machine Translation for Low-Resource Languages: A Surveycited by 109
- 2022A Few Thousand Translations Go a Long Way! Leveraging Pre-trained Models for African News Translationcited by 98
17 more
- 2023The unreasonable effectiveness of few-shot learning for machine translationcited by 91
- 2017Reinforcement Learning for Bandit Neural Machine Translation with Simulated Human Feedbackcited by 72
- 2020Neural Contextual Bandits with Deep Representation and Shallow Explorationcited by 71
- 2018A Survey of Domain Adaptation for Neural Machine Translationcited by 64
- 2022SummaReranker: A Multi-Task Mixture-of-Experts Re-ranking Framework for Abstractive Summarizationcited by 61
- 2021Domain Adaptation and Multi-Domain Adaptation for Neural Machine Translation: A Surveycited by 46
- 2016Domain Control for Neural Machine Translationcited by 36
- 2019Improving Robustness of Machine Translation with Synthetic Noisecited by 36
- 2015A Survey on Contextual Multi-armed Banditscited by 34
- 2019Domain Adaptation of Neural Machine Translation by Lexicon Inductioncited by 33
- 2021Continual Mixed-Language Pre-Training for Extremely Low-Resource Neural Machine Translationcited by 29
- 2022A survey on multi-player banditscited by 19
- 2020Dictionary-based Data Augmentation for Cross-Domain Neural Machine Translationcited by 13
- 2019Adaptation of Machine Translation Models with Back-translated Data using Transductive Data Selection Methodscited by 8
- 2020Incorporating Bilingual Dictionaries for Low Resource Semi-Supervised Neural Machine Translationcited by 8
- 2021Multi-Domain Adaptation in Neural Machine Translation Through Multidimensional Taggingcited by 8
- 2025Faster Machine Translation Ensembling with Reinforcement Learning and Competitive Correctioncited by 1
Built on by
Not cited yet.